Underestimated Formaldehyde Emission from Residential Coal Combustion in North China

Atmospheric formaldehyde (HCHO) plays a critical role in atmospheric radical budgets and secondary formation of ozone and particulate matter. However, HCHO is known to be significantly underestimated by regional air quality models, which indicates an incomplete understanding of HCHO origins and limits our comprehension of its atmospheric roles and implications. In this study, we revealed that direct emissions of HCHO from wintertime residential coal combustion in north China has been significantly underestimated in current emission inventories, based on field measurements. We observed high values of HCHO (up to 9.4 ppbv) at a typical rural site in Qingdao, north China, which exhibited a diurnal variation pattern with a double-peak distribution in the morning and late afternoon. During the morning peak period, HCHO showed a stronger correlation with SO₂ and NO, while HCHO strongly correlated with CO, NO₂ and biomass burning indicators (levoglucosan, K⁺, Cl^-) during the late afternoon peak period. The diurnal variation of HCHO aligns well with the combustion activities of residents, which are dominated by coal combustion in the morning and biomass burning in the evening, implying the potentially significant contributions of residential combustion emissions to ambient HCHO. We then conducted source apportionment using the Positive Matrix Factorization (PMF) method and confirmed the significant contributions of residential biomass burning and residential coal combustion to observed HCHO. Using the Minimum R Squared (MRS) method, we further calculated the HCHO emission ratios from the two combustion-related sources. We further calculated HCHO emission from residential combustion sectors in widely used emission inventories, such as EDGER, MEIC, MIX and REAS. However, HCHO emissions from residential combustion sectors exhibit large discrepancies among these emission inventories, with the maximum estimate provided by MEIC and the minimum estimate by EDGAR. When we calculated HCHO emissions from residential coal combustion sector and residential biomass burning sector separately, HCHO emissions from residential biomass burning sector were well estimated, while HCHO emissions from residential coal combustion sector were significantly underestimated by an order of magnitude. Using current emission inventory, HCHO was significantly underestimated by CMAQ modeling compared with field measurements. After updating the HCHO emissions from residential coal combustion based on field results, modeled HCHO concentrations significantly increased (over 100%), which further enhanced the atmospheric oxidation capacity and secondary organic formation. Our findings highlight the necessity to revisiting the HCHO emission from residential coal combustion sector in current emission inventory, especially in north China.